Connecting Waterpeople

You are here

Predicting water supply and demand: Operating reservoirs using AI

  • Predicting water supply and demand: Operating reservoirs using AI

About the entity

University of Duisburg-Essen
The University of Duisburg-Essen (German: Universität Duisburg-Essen) is a public research university in North Rhine-Westphalia, Germany.

Floods, droughts, heatwaves: Climate change poses increasing challenges for streamflow forecasting and reservoir management. Reservoirs play a crucial role in flood protection and during periods of drought. In order to adapt them to climate change, researchers at the University of Duisburg-Essen are developing a prediction-based control system for reservoirs based on artificial intelligence. The PROWAVE project is being funded by the Bundesstiftung Umwelt with 348,000 euros over the next three years.

The weather events caused by climate change are difficult to foresee and pose a major challenge for the operation of reservoirs and their catchment areas," explains Prof Dr André Niemann from the Institute of Hydraulic Engineering and Water Resources Management at the University of Duisburg-Essen (UDE). "Until now, reservoirs have been managed by a series of river gauges. However, the decisions made on this basis are reactive and slow".

The PROWAVE project is being funded by the Bundesstiftung Umwelt with 348,000 euros over the next three years

In the future, thanks to technological advances, artificial intelligence (AI) methods will support reservoir operators. To this end, special processes such as LSTM (Long Short-Term Memory) networks will enable efficient control strategies. This creates tools that utilise the increasing forecast information in a targeted manner. In combination with modern data analysis tools, they enable proactive control of water distribution systems based on reliable forecasts of water demand and water supply. "Until now, this forecasting was extremely difficult, as every simulated physical process had to be known. Artificial intelligence methods now make the difference," says Gregor Johnen, lead engineer in Prof Niemann's team, and explains further: "For example, neural networks and ensemble methods can use the real-time data and learn from it to simulate complex processes and improve the predictability of extreme weather events through the learned process understanding - up to three months in advance".

The aim of the project is to develop a prototype model for reservoir management and the supply of water for communities and ecosystems. It is being developed using the example of the Harz waterworks' interconnected reservoir compound. This integrates the researchers' predictions into a more complex optimisation model which is provided by the KISTERS AG based in Aachen. "Proactive control of our systems in this way enables us to react flexibly and at an early stage to any future conflicts of use between the water supply and other water users," says Dr Alexander Hutwalker, coordinator of the Research and Development department at Harzwasserwerke.

In the long term, the researchers and their associated partners, the Ruhr- and Wupperverband, would like to use the prototype to achieve the transition to smart reservoir management that meets the requirements of climate change.

Subscribe to our newsletter

Topics of interest

The data provided will be treated by iAgua Conocimiento, SL for the purpose of sending emails with updated information and occasionally on products and / or services of interest. For this we need you to check the following box to grant your consent. Remember that at any time you can exercise your rights of access, rectification and elimination of this data. You can consult all the additional and detailed information about Data Protection.